FINAL PUBLISHABLE REPORT
CONTRACT N° : ENK6-2000-00100
PROJECT N° :
NNE5-1999-20187
ACRONYM :
CEHEG
TITLE : COST-EFFECTIVE GEOPHYSICAL TECHNOLOGY FOR PETROLEUM
EXPLORATION IN NON-SEISMIC AREAS
PROJECT CO-ORDINATOR : UNIVERSITY OF THESSALONIKI (UTHESS.GPL)
PARTNERS : GEOSYSTEM SRL (GEOSYST)
ENTERPRISE OIL EXPLORATION LTD (ENTOIL)
REPORTING PERIOD : FROM 1-9-2000
PROJECT START DATE : 1-9-2000
TO 31-8-2002
DURATION : 24 MONTHS
Project funded by the European Community
under the EESD Programme (1998-2002)
Final Publishable Report
2
CEHEG/ENK6-2000-00100
1
Table of contents
1
2
3
4
Table of contents .......................................................................................................................... 3
Executive publishable summary .................................................................................................. 4
Objectives of the project .............................................................................................................. 5
Scientific, technical description of the results ............................................................................. 6
4.1
WORK PACKAGE 1: Processing Code. ............................................................................. 6
Task 4.1.1. Analysis ................................................................................................................. 6
Task 4.1.2. Implementation...................................................................................................... 7
4.2
WORK PACKAGE 2: Data Acquisition ............................................................................. 9
Task 4.2.1. Planning ................................................................................................................. 9
Task 4.2.1. Fieldwork .............................................................................................................. 9
4.3
WORK PACKAGE 3: Processing Optimisation ............................................................... 11
4.4
WORKPACKAGE 4: Modelling Code ............................................................................. 13
Task 4.4.1 Analysis ................................................................................................................ 13
Task 4.4.2 Implementation..................................................................................................... 13
4.5
WORKPACKAGE 5: Interpretation .................................................................................. 15
4.6
WORKPACKAGE 6: Reporting ....................................................................................... 17
Task 4.6.1 Results .................................................................................................................. 17
Task 4.6.2 Exploitation .......................................................................................................... 19
5
List of deliverables ..................................................................................................................... 21
6
Scientific and technical performance ......................................................................................... 22
6.1
Comparison of initially planned activities and actual work accomplished. ....................... 22
6.2
State of the art review: comparison of the project current work with the competing
technologies and assessment of its implications on the relevance of the project. ......................... 22
7
Management and co-ordination aspects ..................................................................................... 24
8
Results and Conclusions ............................................................................................................ 29
9
Acknowledgments ...................................................................................................................... 30
10
Glossary ................................................................................................................................. 31
11
References .............................................................................................................................. 32
12
Annex (Individual Reports for each Work Package as Provided by the Work Package
Leader) ............................................................................................................................................... 34
Final Publishable Report
3
CEHEG/ENK6-2000-00100
2
Executive publishable summary
The Magnetotelluric (MT) method is a maturing geophysical method. It is mostly used in projects
for deep structure (up to 8 Km for oil exploration and 100 Km in academic research). Its advantage
compared to other geophysical methods is that it is very cost effective. However, as a newly
developed method (just fifty years of its first use) it is necessary to evolve the processing
techniques especially since the method is sensitive to man-made noise sources. Also, on
interpretation it is imperative to improve the sub-surface imaging techniques applicable to nearvertical structures with a strong degree of three-dimensionality.
An area in NW Greece has been selected as a test area and field measurements of MT and Time
Domain Electro-Magnetic (TDEM) has been acquired. The area consist of an oil target in difficult
terrane where other geophysical data exist and which will be used to verify the results of the
electromagnetic (EM) methods applied in the area.
During this project a processing method based on the remote reference technique has been
developed and implemented into software that has been tested and optimised through the first year
of the project. The data acquired during the project were used in order to finalise the processing
code implemented.
Analysis and implementation of an inversion code to be used for the interpretation of the MT data
has been established. The code, able to produce 3-D imaging of the earth, has been tested into
synthetic data and has also been used for modelling the MT data acquired. This included
incorporating constraints mostly from wells and less from other available geophysical data. Based
on the contrast of the physical properties of the strata, mapping of the reservoir was possible. This
results in implementing the state of the art inversion code in two and three-dimensional modelling
of magnetotelluric data.
The main project results could be summarised in:
1) Processing code of magnetotelluric data
2) Data set of magnetotelluric time series
3) Two and three dimension modelling code of magnetotelluric data
4) Geophysical/Geological models of the test area
5) Environmental/Economic Impact of the application of the magnetotelluric method in
hydrocarbon exploration
The above-mentioned results are useful since we managed to:
1) Reduce the time necessary to evaluate the dimensions of the hydrocarbon reservoir;
2) Validate the proposed geophysical method in exploiting difficult geological structures;
3) Transfer of knowledge to an end-user in the petroleum industry
Even before the end of the project and ever after the results of the project are and will be fully used
from all the partners. In more details, the geophysical company involved (Geosystem Srl) is
making use of the processing and modelling codes implemented. That will enhance its ability to
manage in less time big magnetotelluric projects. The hydrocarbon company involved (Enterprise
Oil Ltd), has already make use of the geophysical/geological models of the test area for further
research or planning of use of the area potential. Also the environmental/economic impact of using
the magnetotelluric method it is of great importance to the oil company since in this way a more
environment friendly way for exploration is introduced and also the expenses for exploration could
be extremely decreased in such a difficult environment. The University involved (Aristotele
University of Thessaloniki) had the opportunity to be involved in the acquisition of a big
geophysical survey and work together with field geophysicists and a production exploration team.
Also the know how transferred from both private companied to the research team of the University
is of great importance since it is a starting point for the further probe in the field of exploration
geophysics.
__________________
Final Publishable Report
4
CEHEG/ENK6-2000-00100
3
Objectives of the project
Magnetotellurics, until about ten years ago, was an underdeveloped geophysical method, due to the
lack of adequate processing and interpretation algorithms. This lead to under-use of the method, but in
the last few years there has been an increase in its application to various geological problems. For this
reason there is a need for developing and implementing different approaches for solving processing
and interpretation problems, including the exploration of hydrocarbon targets.
The main objectives of the project is to extend the exist technology of the application of the
magnetotelluric (MT) sounding method into areas where the structure are very complex, such as a
hydrocarbon target. Also to apply this geophysical method in areas where acquisition, processing and
modelling of magnetotelluric data present significant problems. Difficulties in acquisition and
processing are due to noise that is of man-made origin. Since there is nothing to be done at the level of
data acquisition systems, all our effort has been concentrated into advance processing techniques,
which will function efficiently in areas of high industrial noise. A state of the art multi-station robust
remote reference processing code has been implemented for calculating the transfer function of MT
time series data. A set of 99 MT soundings has been acquired and the data has been used to test,
optimise and evaluate the processing code.
Concerning modelling of the MT sounding data a 3-D imaging inversion modelling code has been
implemented. Already a minimization of time has been achieved to the first computed model of the
area of study. Based on these geophysical models and using constrains and the physical properties of
the geological formations mapped in the area of study a geological interpretation is presented.
Finally the usage of the MT method successfully managed to transform geophysical exploration into a
non-destructive and cost-effective approach. By increasing the applications of the magnetotelluric
geophysical method for hydrocarbon exploration more information will be available to the oil
industry regarding the cost-effective application of this technique in difficult geological and/or
noisy areas. To accomplish this we have been working in three main areas:
a) high resolution processing of MT;
b) 2-D and 3-D MT modelling and/or inversion; and
c) evaluation of an MT test survey.
The results of these three areas of research will guide hydrocarbon exploration towards cost
effective and timesaving exploration methodology, which is particularly relevant to large sections
of the European Union where conventional techniques fail, but where geological conditions may be
attractive.
The work already done contributes to increasing the commercial success rate of exploration,
minimizing both time and cost. Also, the geophysical method applied has no negative
environmental impact, since the natural electromagnetic field is the source for the evaluation of the
earth structure. This will avoid the usage of high volumes of explosives that are necessary for the
seismic geophysical method.
The cost of the project is much less (by at least a factor of thirty) than the cost of a seismic
exploration project. That indicates that the results obtained through the project is of significant
economic value and thus the adoption of the magnetotelluric method as an exploration tool for
hydrocarbons show a significant improvement in the benefit/cost ratio with respect to currentlyused techniques.
Final Publishable Report
5
CEHEG/ENK6-2000-00100
4
4.1
Scientific, technical description of the results
WORK PACKAGE 1: Processing Code.
Task 4.1.1. Analysis
Analysis of problem
Fundamentally, the problems to be resolved encompass
 Estimation of impedances in presence of very high levels of both harmonic and transient noise
 Accomplishment of this on a time-scale sufficiently short to be of practical use in the field.
Thus, the tasks to be addressed can be broken down into two phases:
1. Identification and implementation of the “best” processing technique;
2. Acceleration of this to reasonable speed with commonly available field computers (typically
PC’s using a Windows OS).
Noise sources include, power lines, cathodically protected pipelines, DC/AC railroads and a
combination of all these that is very common in industrial cities.
Evaluation of existing codes
Codes for the processing of MT time-series utilise some of the following techniques:
1. Spectral estimation through FFT (Fast Fourier Transform) or DFT (Discrete Fourier
Transform) techniques;
2. Remote referencing to distant and presumed quiet magnetic reference sites;
3. Robust estimation of spectral parameters;
4. Pre-treatment (de-spiking, filtering, etc) of time series.
In regard to the last, we implemented a despiking routine based on the wavelet transform (Trad and
Travassos, 2000). This did not prove to be useful; since the underlying assumption is that (in time)
the noise spikes are instantaneous. In reality, they have significant time dependence, causing the
method to be of little use when applied either in time or frequency domains.
In this project, we evaluated several published techniques for estimating magnetotelluric transfer
functions: Chave (Chave and Thomson, 1989), Egbert (Egbert 1997), Cascade decimation (Wight
and Bostick, 1981; Jones et al., 1989), Larsen (Larsen et al., 1996).
The methods referenced above use the following approaches:
Cascade
decimation
Chave
DFT, decimation of time series coherency sorting;
Egbert
Standard FFT to estimated binned parameters followed by
robust multivariate analysis;
Larsen
Time-and frequency domain despiking, signal separation, robust
coherency estimation of impedances;
Final Publishable Report
Standard FFT to estimate single-frequency parameters,
computation of coherency between local and remote sites, and
rejection of outliers (presumed contaminated by noise) using
advanced robust statistical procedures;
6
CEHEG/ENK6-2000-00100
Task 4.1.2. Implementation
The algorithms listed above were implemented using FORTRAN 95 and compiled to run under
Microsoft Windows operating system.
Testing of code
The codes have been tested on two data sets obtained in the Southern Apennines of Italy, referenced
to a magnetically-quiet site 300 Km further south east in Calabria. This ensures that processing will
not enhance coherent noise between the field site and the reference station. Two separate data sets
have been used, one from a fairly quiet site, the other from a much noisier site in the same general
area, each recorded for about 13h.
Testing on data from quiet site
The field data were recorded from 1530 to 0417h at 64Hz sampling rate, for a total time of nearly
13 hours and 11.5 MB of data per channel. At the higher sampling rate, 512 Hz, two runs recorded
at 0430 and 0530h were used, each lasting 512 seconds and resulting in a 1.1 MB file per channel.
Two reference sites were available, located a few hundred meters apart, each recording the two
horizontal magnetic channels.
Table 1: Processing times with different algorithms: quiet site
Processing
Type
Cascade
decimation
Chave
Egbert
Larsen
Description
Local H -Remote
Remote
reference.
Robust
processing
Remote reference. Two step
processing.
As option 1, but with 2 remote
references.
Robust processing with 2 remote
references.
Signal
separation
LS-LS.
Decimation: 1.
Remote Reference. Decimation: 2.
FFT length
Sample freq
[Hz]
n/a
512-16-1
4096
512-64-10.0625
512-64-10.0625
512-64-10.0625
512-64-16-10.25-0.00-625
512-64-2-0.125
00:05:00
1-0.25-0.0625
00:52:40
4096
4096
128
2048
512 4096
8192
Processing
time
(hh:mm:ss
)
00:01:11
00:05:40
00:12:50
00:13:34
00:37:50
Notes: (1) Processing times are for a 233MHz Pentium II PC with 500MB of RAM.
(2) Data at frequencies below 64Hz obtained by decimation from 64Hz data.
The data from this site present only minor problems to the different processing algorithms, with the
exception of the Egbert code. Harmonic noise at 50Hz, and its odd and even harmonics is up to
nearly three orders of magnitude greater than the natural MT signal. This biases the output from all
programs except Larsen (which automatically identifies and removes spectral lines), and these and
other outliers have been removed from all plots. In the case of the Egbert algorithm, harmonic noise
rendered the computational process so unstable that no results were produced at nearby frequencies.
Overall, the results are similar for each method (with the exception of Egbert), with Larsen showing
somewhat less scatter at longer periods. In terms of computational time, and using a single remote
reference site, Larsen is clearly the most inefficient (Table 1), requiring about 18 times more CPU
time than the other methods.
Testing on data from noisy site
The noisy test site is located on high resistivity carbonates, about 20 Km from the quiet one. The
effect of the high resistivity earth is to minimise the attenuation of electromagnetic noise. In this
particular case, as well as harmonic noise, transient noise sources are associated with cathodically
Final Publishable Report
7
CEHEG/ENK6-2000-00100
protected pipelines and with a major DC electric railway. The noise continues night and day, i.e.
there are no windows of time greater than about 15 seconds when noise is not evident.
Table 2: Processing times with different algorithms: noisy site
Processing Type
Description
FFT length Sample freq
[Hz]
Processing
time
(hh:mm:ss)
00:02:32
Cascade decimation
Local H -Remote.
n/a
512-64-1
Chave
Remote reference.
4096
512-64-10.0625
512-64-10.0625
512
00:12:03
64
00:57:24
64
03:18:09
1-0.250.0625
00:10:02
Larsen
Remote
reference.
2
step
4096
processing
Signal
separation
LS-LS.
2048
Decimation: 1.
Remote Reference – Time series
1024
skipped at 20'000sec.
Signal
Separation
RR-LS.
2048
Decimation: 1.
Remote Reference. Decimation: 1. 512-40968192
00:15:43
00:15:14
In this test, the Egbert code produced no usable data, and is not therefore further discussed. With the
exception of the Larsen code, the processed data pass from far field to near field at a period of about
2 seconds. This is evidenced by a sharp rise in both apparent resistivity and tipper, and an
accompanying decrease of the impedance phase to zero. Cascade decimation also appeared to be
more strongly affected by power line noise, resulting in degraded parameter estimates at higher
frequencies.
The Larsen algorithm resulted in unusable data with Options 3 and 5, and so the results shown
above are from the two stable methods, Options 2 and 4. It is clear that the results are less strongly
affected by noise than the other techniques, but data quality at periods greater than about 10s
remains poor.
Conclusions and Innovations
In terms of the efficiency of computation, and minimization of bias by noise, the Chave code is the
most attractive under conditions of low to moderate noise. At higher noise levels, the only code
which produces magnetotelluric parameters from which noise appears to have been successfully
removed is the Larsen code. Unfortunately this requires a substantially greater amount of CPU time,
to which should be added greater time needed by a geophysicist for the assessment of results.
Nevertheless, these two codes (Chave and Larsen) offer the best solution to the computation of
magnetotelluric parameters.
The innovations succeeded through this work package can be concluded into the following:
 For the first time all the algorithms referenced above have been implemented and tested
together.
 A user friendly tool has been implemented so that the geophysicist dealing with the processing
of the data can easily and quickly evaluate the quality of the time series and calculate the
sounding data.
 A review of the ability of different processing techniques to overcome noise problems is
presented and an estimate of processing time is given depending on algorithm used and noise
level.
Final Publishable Report
8
CEHEG/ENK6-2000-00100
4.2
WORK PACKAGE 2: Data Acquisition
Task 4.2.1. Planning
Internal reports, well data and digital geological and geophysical data were available to the
consortium from Enterprise Oil Ltd. Also, other bibliographic information was collected concerning
the tectonic and geological setting of an area broader than the one to be studied. Field crews visited
the area of Filliates and Dholiana, where evaporates outcrop, in order to check the geology in more
detail and the possibility of establishing a 10 by 10 km grid in both sites, as proposed. The
possibility of working in the area of Dholiana arose after the evaluation of the geophysical and well
data. In more detail, for the area of Filliates, there are data from one borehole only. In the area of
Dholiana, not only data from four wells exist, but also there were available gravity, magnetic, some
magnetotelluric records and seismic tomography data. Based on this information, we decided to
establish our grid in the area of Dholiana.
In order to define the specific area of study, in a convenient manner, all digital data where imported
and processed through Geographical Information System (GIS) software. A digital geological map
is presented in figure 1 which was produced by the GIS software. Different geological formations
are given in appropriate colours. Previous MT stations are depicted with red circles, and the
position of the wells with yellow ones. The seismic tomography grid is presented with a blue polyline and a yellow line show one of the seismic lines from which data were available to the
consortium. All the geological and geophysical information were evaluated in order to define the
area of study that is denoted with a red polygon in figure 1.
Task 4.2.1. Fieldwork
The fieldwork has been accomplished in March to May 2001, and included 99 magnetotelluric
(MT) soundings on a grid of 9 lines (NW-SE direction) x 11 rows (SW-NE direction). The grid
distance was 1.35 km and at the same positions time-domain electromagnetic (TDEM) soundings
were carried out for the purpose of effecting static shift corrections.
Prior to starting the survey all the coils were tested. The remote reference technique was used
during the MT data acquisition. After the first recordings at the start of the project it was clear that
the southeast part of the survey area was noisier than the rest of the grid. Based on this information
we planned everyday to have two stations away from this sub-area in order to use them as remote
references. Stations were recorded overnight for a total of about 14h per site, and through the use of
multiple systems (2 operators deploying up to 4 systems per day) a high acquisition rate (4.5
stations/per day ) was achieved, leading to a more cost-effective survey.
The data were archived daily and QC procedures were established in the field and in the office
before the data processing. The time series were processed overnight using the newly developed
algorithms implemented in Work Package 1 (WP1). In the morning after the evaluation of the
processed data the following sites were planned (including repeats if necessary) and the crews left
the office for the field.
The TDEM data were acquired at each MT site using 100 m x 100 m square loops. The maximum
time range was approximately between 0.08 to 10 ms, providing adequate overlap with the MT
curves for static shift purposes.
Final Publishable Report
9
CEHEG/ENK6-2000-00100
Figure 1. Geology map where the positions of prior MT soundings are denoted with red circles and the positions
of wells with yellow circles. The area enclosed in the blue box is the area covered by the tomography results. The
yellow line corresponds to the seismic line ENT-E97GR09-01. The area denoted with the red polygon is the area
of the EM survey and the area with the magenta color is the polygon of the 3D MT inversion applied.
Final Publishable Report
10
CEHEG/ENK6-2000-00100
4.3
WORK PACKAGE 3: Processing Optimisation
The codes previously described in WP 1 have been optimised to run under Windows NT and 2000
operating systems. The efficiency of the code has been improved by the use of dynamic memory
allocation, which speeds up processing if data blocks are less than the previously assigned fixed
array sizes. The selection of data and the settings required for each processing technique are
accomplished through a user interface.
The robust code was used to process all of the data collected in the Delvinaki test area. Initially, the
Chave code was used to check data quality. This has the advantage of being fast but is not
particularly good at eliminating noise. In the event that problems were noted at this point, the
sounding was repeated. Data that seemed to be correct, that is, no evidence was seen of equipment
malfunction or insurmountable noise sources, were then processed using the Larsen code. In general
the remote-reference signal separation option was used, the reference site being selected on the
basis of the preliminary processing.
Figure 2. Time series from station 05E07N, sampling rate 512Hz. Strong harmonic noise at multiple frequencies
is visible on all channels.
Noise sources in the Delvinaki area fell into two categories: strong harmonic noise associated
mostly with power lines, and low frequency square waves probably caused by cathodic protection
circuits on water pipelines. An example of the former is shown as time series in figure 2.
Transformed to the frequency domain, it is clear from the spectrum of figure 3 that the noise is due
to 50Hz and both its odd and even harmonics.
An example of the second type of noise that attributed to cathodic protection circuits is shown in
figure 4. In this, square waves with a period of about 50 seconds are clearly visible and in addition
to this a major disturbance is visible near the beginning of this example. The frequency domain
transform of these data is a broad series of spikes, and notch filtering suitable for harmonic noise, is
of little use in this case. Instead, the Larsen algorithm uses the transfer functions from a quiet site to
eliminate as far as possible spikes and jumps in the time domain. That this was reasonably
successful is shown by the examples below.
Final Publishable Report
11
CEHEG/ENK6-2000-00100
Figure 3. Spectrum of first channel (Ex) of above time series; arrow shows 50Hz peak. Broader peaks closer to
10Hz are Schumann resonances.
Figure 4. 4Hz time series from site 06E01N showing 50s square waves on Ex.
The soundings within the test area fall into two classes. Figure 5 shows a typical sounding made on
outcropping carbonates or anhydrite, with the crossover of apparent resistivity curves at about 30
Hz indicating 3D effects. The second type (figure 6) corresponds to soundings made on outcropping
(conductive) flysch. Noise was rarely a problem in the latter case, since the conductor effectively
screened out much noise. However, soundings made on the higher resistivity carbonates were more
subject to noise problems, the examples referred to the above figures 3 and 4 being typical.
However, the Larsen code even in this case gave acceptable results (figure 7).
Figure 5. MT sounding 09E07N.
Final Publishable Report
Figure 6. MT sounding 02E07N.
12
CEHEG/ENK6-2000-00100
Figure 7. MT sounding 06E01N.
4.4
WORKPACKAGE 4: Modelling Code
Task 4.4.1 Analysis
2D Magnetotelluric Sharp Boundary Inversion
The Gauss-Newton algorithm has been analysed and finally utilized for the 2D sharp boundary
inversion. It is used to minimize an objective function that is the sum of the squared data misfits
normalized by their variances and a model norm which is related to the gradients of the resistivity
within each layer plus the gradients of the vertical position of the nodes of each interface. Thus, the
algorithm will try to produce the flattest interfaces and flattest resistivity structure within the layers
consistent with the data. The user can specify the trade off between data misfits and the model
norm.
3D Magnetotelluric Inversion
For the 3D magnetotelluric inversion the algorithm is based on the non-linear conjugate gradient
(NLCG) method to minimize an objective function that is the sum of the squared data misfits
normalized by their variances and the model norm defined by an approximation to the Laplacian for
the model grid. What makes the NLCG method so useful for an inverse problem is the fact that it is
not necessary to compute or store the Jacobian (sensitivity matrix) or the Hessian of the objective
function. Rather, it is only necessary to compute the result of those matrices multiplying arbitrary
vectors.
Task 4.4.2 Implementation
2D Magnetotelluric Sharp Boundary Inversion
A 2D magnetotelluric (MT) sharp boundary inversion algorithm has been developed that inverts
MT data for a 2D model consisting of discrete interfaces and the resistivities of the layers between
those interfaces. The interfaces are described by a series of nodes, whose horizontal positions are
fixed, but whose vertical positions can vary in the inversion. The interfaces are assumed to transect
the entire model, i.e., there are no closed bodies. The interface varies linearly between each
interface node. The resistivity of each layer is also described by a set of nodes at fixed horizontal
positions within each layer. The resistivity is assumed to vary linearly between nodes. The interface
and resistivity information is projected onto a finite-difference mesh for computation, and the
inversion calculates the best fitting interface node locations and resistivity nodal values in order to
fit the observed data.
Final Publishable Report
13
CEHEG/ENK6-2000-00100
3D Magnetotelluric Inversion
A 3D magnetotelluric (MT) inversion algorithm that is based on previous work of GEOSYST’s
personnel for 2D MT inversion (Rodi and Mackie, 2001) and 3D MT forward modeling (Mackie et
al., 1994) has been developed. As it is presented in Rodi and Mackie (2001), the Jacobian
multiplying a vector could be performed with only one pseudo-forward problem requiring the same
amount of time as the actual forward problem. For the NLCG method, it turns out that 3 forward
problems per inversion iteration are required. It should be noted that there are no approximations
made in this code except for the finite difference solution of the 3D grid. The net result is that we
are able to invert 3D data (each element of the impedance) for smooth 3D models. As an example,
we have inverted real data comprising 95 stations and 5 frequencies using a model grid of 31x31x21
(not including air layers). An inversion of 20 iterations took approximately 10 hours of CPU time
on a 400 Mhz computer running the Linux operating system. In figure 8 the preliminary results
from the application of the implemented code using the data acquired through this survey and
previously collected data for ENTOIL, is shown.
Current
Survey
Line A
Line B
Approx position of new well
Figure 8. Preliminary results from 3D inversion. Top panel shows resistivity at 3000m below ground level
(red=resistive, blue=conductive). Black triangles are positions of MT stations. Lines A and B are from an older
survey. White lines show possible faults. Lower panel shows section through 3D model along line B of an older
survey in the area. 2D inversion is basically similar, i.e. Resistive to about 4000m, then more conductive.
Final Publishable Report
14
CEHEG/ENK6-2000-00100
4.5 WORKPACKAGE 5: Interpretation
In this work-package an attempt is presented to interpret the different geophysical data available to
the consortium. From these data only the Passive Tomography data acquired for EntOil in the past
and the current acquired EM (electromagnetic) data were of practical use in the construction of
geophysical models that could be interpreted. The interpretation of the Passive Tomography data is
not presented in this report since they are confidential to the consortium and not publishable.
The MT data used comprise of 99 soundings on carbonates, Triassic evaporites and the flysch as
presented in Figure 9.
TR
Ev
a
Ca
rb
po
rit
es
on
ate
s
Li
ne
07
N
h
c
ys
Fl
Li
ne
11
N
Delvinaki-1
Li
ne
03
N
Lavdani-1
Figure 9. Geological Map with the EM sounding positions and the wells used for constrains.
The code implemented through WP4 was used for modelling the data acquired. Both 2D sharp
boundary inversion and 3D inversion was applied to the data. A 2D model from the sharp boundary
inversion is presented in WP4. In this work package we concentrated on the results from the 3D
inversion since it better images the interfaces between different formations.
For the 3D inversion we constructed a model of 47x45x26 blocks, on x, y, z, directions
respectively, (plus 10 air layers), with minimum distance of 680m in the horizontal plane and 30m
in the vertical dimension. We used 119 sites, the 99 sites of this MT project and 20 sites (sections A
and B) of a previous MT project of ENTOIL. Eleven equally spaced frequencies from 100 to 0.001
Hz were used from all the components of the impedance tensor assuming a 5% error floor. The
starting model consisted of a half-space of 25 ohm-m. An rms misfit equal to 2.0 was accomplished
after 20 iterations.
Various 3D videos were created for the resistivity data resulted from the inversion of the MT data
and one is presented below. The video shows the distribution of the logarithm of the resistivity as it
was obtained by the 3D inversion. The colour scale starts from the magenta and blue colours for the
lowest values to reach the red colour with the highest resistivities passing through green and yellow.
Final Publishable Report
15
CEHEG/ENK6-2000-00100
Please double click on the image on the left to play the video
Playing the video, one can see the position of the geological formation of flysch that is presented
with the blue/magenta colours (at the last part of the video) due to its correspondence to low
resistivity values.
It seemed more useful to work with cross sections of the 3D models, since they were easier to
visualize and interpret one by one.
In Figure 10 the physical properties (resistivity) of the geological formations are shown in the form
of a stratigraphic column.
Flysch
<20 Ohm.m
Carbonates
200-2000 Ohm.m
Evaporites
200-10000 Ohm.m
?Pz Clastics
?? 30-100 Ohm.m
?pC Metamorphic
basement
Figure 10. Geological formations of the area of study and their physical properties (resistivity).
Based on this information and the tectonic setting, we proceeded with the interpretation of the
cross-sections from the 3D models. In Figure 11 the interpreted model of Line 07 is presented.
Final Publishable Report
16
CEHEG/ENK6-2000-00100
Flysch
Car
Car
Evaporites
Car
??
Evaporites
??
??
Figure 11. Interpretation of Line 07N in terms of lithology.
The base of the Triassic evaporites is presented in Figure 12.
Figure 12. Elevation of base Triassic evaporites (in meters asl).
4.6
WORKPACKAGE 6: Reporting
Task 4.6.1 Results
(a) Processing Code of Magnetotelluric Data
Published algorithms were implemented and used to process data collected in Northern Greece.
Broadly speaking, the approach were based on the use of coherency sorting and/or spectral despiking after estimating cross-spectra with FFT or DFT techniques. The most successful results, in
conditions of moderate industrial noise, were obtained using an adaptation of an algorithm
published by Larsen et al., (1996). Through the use of advanced Fortran programming techniques,
processing time for a single site with a reference site, each with about 600MB of recorded data, was
reduced from 24h of CPU time to about 6h on a Pentium IV computer. The resultant code has been
Final Publishable Report
17
CEHEG/ENK6-2000-00100
introduced into the commercial processing chain, resulting in faster data turn-around with a clear
advantage as far as in-field QC is concerned. Poor data quality can be rapidly identified and the
acquisition process repeated, or the site re-located, whilst the field crew is still in the survey area.
(b) Data set of Magnetotelluric Time Series
The magnetotelluric (MT) data set consists of two types of data; a) the time series of the recorded
electric and magnetic field in 99 sites that follow a grid of 9 rows by 11 columns, where its grid
point is at a distance of approximately 1.3 km from the adjacent one, and b) the transfer functions
connecting the electric and magnetic field at each site as a result of the processing code
implemented through this project (result a). Both data sets are on the deliverables of the project and
belong to all the partners of the project. It is the first time in Europe that such a big amount of MT
soundings exist in one area following a grid of approximately 100 square kilometres, that allow two
and three-dimensional modelling.
(c) Two and Three Dimension Modelling Code of Magnetotelluric Data
We have developed a 2D magnetotelluric (MT) sharp boundary inversion algorithm that inverts MT
data for a 2D model consisting of discrete interfaces and the resistivities of the layers between those
interfaces. The interfaces are described by a series of nodes, whose horizontal positions are fixed,
but whose vertical positions can vary in the inversion. The interfaces are assumed to transect the
entire model, i.e., there are no closed bodies. The interface varies linearly between each interface
node. The resistivity of each layer is also described by a set of nodes at fixed horizontal positions
within each layer. The resistivity is assumed to vary linearly between nodes. The interface and
resistivity information is projected onto a finite-difference mesh for computation, and the inversion
calculates the best fitting interface node locations and resistivity nodal values in order to fit the
observed data. We use the Gauss-Newton algorithm to minimize an objective function that is the
sum of the squared data misfits normalized by their variances and a model norm which is related to
the gradients of the resistivity within each layer plus the gradients of the vertical position of the
nodes of each interface. Thus, the algorithm will try to produce the flattest interfaces and flattest
resistivity structure within the layers consistent with the data. The user can specify the trade-off
between data misfits and the model norm. The starting point for such inversions is often the 3D
inversion (below), constrained by available borehole data.
We have developed a 3D magnetotelluric (MT) inversion algorithm that is based on our previous
work for 2D MT inversion (Rodi and Mackie, 2001) and 3D MT forward modelling (Mackie et al.,
1994). The inversion algorithm is based on the non-linear conjugate gradient (NLCG) method to
minimize an objective function that is the sum of the squared data misfits normalized by their
variances and the model norm defined by an approximation to the Laplacian for the model grid.
What makes the NLCG method so useful for an inverse problem is the fact that it is not necessary to
compute or store the Jacobian (sensitivity matrix) or the Hessian of the objective function. Rather, it
is only necessary to compute the result of those matrices multiplying arbitrary vectors. We showed
in Rodi and Mackie (2001) how, for example, the Jacobian multiplying a vector could be performed
with only one pseudo-forward problem requiring the same amount of time as the actual forward
problem. For the NLCG method, it turns out that 3 forward problems per inversion iteration are
required. It should be noted that there are no approximations made in this code except for the finite
difference solution of the 3D grid. The net result is that we are able to invert 3D data (each element
of the impedance) for smooth 3D models. As an example, we have inverted real data comprising 95
stations and 5 frequencies using a model grid of 31x31x21 (not including air layers). An inversion
of 20 iterations took approximately 10 hours of CPU time on a 400 Mhz computer running the
Linux operating system. The code has now (1/2003) been used on more than a hundred separate
data sets. Its advantages include (1) use of all processed data (impedances), including on-diagonal
terms for all sites and most frequencies for a given survey (2) elimination of need to force data into
an inappropriate 2D framework (3) the computation of a reliable, consistent, 3D resistivity model
which can then be interpreted in terms of deep geology. The use of the code in the industrial
Final Publishable Report
18
CEHEG/ENK6-2000-00100
environment has been growing, commercialised as a service by one of the partners (Geosystem). It
has been applied to oil and other resource projects in Albania, Bolivia, Canada, Chile, El Salvador,
Greece, Iran, Italy, New Zealand, and the USA.
(d) Geophysical and Geological Models
The regional part of the area of study is of exploration interest of one of the partners, i.e. SHELL,
who is the end user of this result. For this reason a lot of geological/geophysical data were collected
for this area. Except from the magnetotelluric (MT) data acquired other geophysical data, as,
gravity, magnetic, passive tomography data, and one seismic line were available to the consortium
from the oil company. From these data sets the passive tomography results and the seismic line
were of interest. Also a geological map and well logs were available. The well logs were used for
the modelling and interpretation of the geophysical data. Using the modelling code implemented
from Geosystem Srl., we came up with some geophysical models. The MT data were the input to
the two and three-dimension inversion code and we came up with different two dimensional
geoelectrical models and one three dimensional model. Also the passive tomography results were
used as these have been modelled from the personnel of the oil company, and we came up with a
three-dimensional model of Vp and Vs velocities. Both three-dimensional models were interpreted
individually and comparisons were applied into selected two-dimensional slices of each cube
between the geoelectrical model and the velocity models. The resulted interpretation suggests the
position of the known geological formations (evaporites, clastics and carbonates) with depth.
(e) Environmental and Economic Impact of the Application of the MT Method in Hydrocarbon
Exploration
Exploration teams in hydrocarbon exploration have been using the seismic geophysical method for
more than seventy years. Explosives are mainly used for this method as an active source. In that
case there is a negative environmental impact caused both by the explosion and by the drilling of
the shot hole. In addition, access of drill rigs often requires the use of bulldozed tracks, which lead
to increased erosion and to unsightly scars on the landscape. On the other hand the use of the
magnetotelluric method is completely friendly to the environment. There is no active source since
the natural electromagnetic field of the earth is the main source. The sensors are recording MT
signals that are due to induction currents produced because of the different geoelectrical properties
of the geological formations. Equipment and personnel arrive at the measurement points on foot, so
no damage to the landscape occurs. In this ways there is not any environmental impact and the area
of study is as before the measurements.
Also, the seismic method is of very high cost, especially in areas of difficult topography as in the
area of study. The typical cost of acquisition in this environment is 27000 EURO per kilometre.
Using this cost in order to cover the area of study an amount of 3.375MEURO would be necessary.
On the other hand for the fieldwork applying the MT method only 0.110MEURO was necessary. Of
course one can say that the resolution of the two methods in imaging the earth might not be the
same. In this specific case, we have shown that the resolution of the MT method has been increased
especially with the processing and modelling techniques developed and implemented through this
project, whilst the seismic surveys in the same area produced essentially no usable data. The points
above therefore show that there is a significant cost/benefit improvement by using the MT method
in hydrocarbon exploration in difficult environments.
Task 4.6.2 Exploitation
Exploitation of the results was direct and straightforward since in the consortium two private
companies were included. At first, the software for the processing code and the two and three
dimension modelling code was the main return directly used into Geosystem Srl’s projects. The
UTHESS.GPL exploited from the software implemented since their personnel had the opportunity
Final Publishable Report
19
CEHEG/ENK6-2000-00100
to work with such a code, got acquainted with it, and got the advantage of using a tool for
geophysical exploration.
The data set of MT data was mainly exploited from UTHESS.GPL and Shell since the first
contractor is having the opportunity to work and practise with a unique data set, and the second one
is having in the test area that is of their rights a denser imaging of the earth.
The geophysical and geological models of the test area were mainly an advantage of Shell since this
information was utilized directly, at the time of the fieldwork. This notice was of big advantage to
Shell since they have used this data at the same time on the drilling for the well plan. Also,
UTHESS.GPL used the data set and will have a future advantage of it.
Finally, the environmental and economic impact of the use of the MT method in hydrocarbon
exploration is a big advantage mainly for Shell since they can use a cost effective geophysical
method with an environmental impact. This advantage will be exploited continuously from the
contractor in the future since Shell is working on exploration projects all over the world. Finally the
exploitation of the MT method is transferred also in other energy problems as geothermal
exploration. The MT method is routinely used as the leading method in geothermal exploration. The
results of this project will give another advantage in the MT method and a perspective into
geothermal renewable energy resources.
Exploitation of the first result, i.e. processing code of magnetotelluric data, has started before the
end of the first year of the project since Dr Savvaidis visited the University of Uppsala and
presented the resulted software to the electromagnetic induction group of geophysicists. Their
comments were welcome and used to further develop the software.
Two publications came out from the project. Both were presented in the EAGE 2002 conference
held in Florence, Italy. The full reference is given below:
1. Watts D., Savvaidis A., Karageorgi E., and Mackie R., Magnetotellurics applied to subthrust petroleum exploration in Northern Greece.
2. Karageorgi E., Watt D., and Savvaidis A., Integration and correlation of geophysical data in
NW Greece.
A web site at http://lemnos.geo.auth.gr/CEHEG/ exists for the project where information for the
project Work Packages and their reports can be found in order to have the geophysical and
petroleum community informed about the results of the project.
Final Publishable Report
20
CEHEG/ENK6-2000-00100
5
List of deliverables
All of the deliverables has been implemented as planned in the deliverable list (Table B2 –
Deliverable list) initialised through the Contract Preparation procedure and are included in the
“Description of Work”. In more details:
D1. Preliminary processing algorithms, delivered at the forth month of the project as planned. The
processing algorithms where analysed and implemented into a software tool, with a Graphical
User Interface (GUI) under Microsoft Windows Operating System (O/S). The software was
tested using time series from previously accomplished data sets. A detailed report from the
work package (WP1) leader (GEOSYST) describing the robust processing algorithms used, the
code implemented and the results from test computations was available after the end of the
work package.
D2. Field test data, available to the consortium after the eighth month of the project as planned. All
geological and geophysical data available to the consortium by ENTOIL are presented by the
project leader (UTHESS.GPL) in both electronic and paper form. The electronic data are
presented through a Geographical Information System (software). A data set of time series,
which include MT and TDEM soundings at 99 stations that follow a grid of 9 rows by 11
columns are available on CD-ROM. A detailed report describing the data available and the data
accomplished as well as the procedures used to collect the data are available to the consortium
from the work package (WP2) leader (UTHESS.GPL).
D3. Completed processing algorithms are the final result of the analysis and implementation of the
available and newly developed code. A fully tested code of multi-station robust remote
reference processing of MT time series data has been implemented. A final report on the final
code is available to the consortium covering the above described deliverable by the work
package (WP3) leader (GEOSYST).
D4. Completed modelling algorithms are fully available to the consortium. The code has been
partially implemented and first results were available to the consortium before the end of the
first year. However, more work had been done by the work package (WP4) leader (GEOSYST)
after the first year in order to optimise and finalise the modelling code. The code was finalised
by the fifteenth month of the project causing a delay on the deliverable for three months. A
description of the modelling code is available to the consortium from the work package (WP4)
leader (GEOSYST).
D5. Process test data has been finalised by the end of the first year and the resulted sounding curves
are available and already been used for testing the modelling code. A description of the final
apparent resistivity curves is presented in the report from the work package (WP3) leader
(GEOSYST).
D6.Interpretation of test data & other geophysical data are available to the consortium four
months before the end of the project. Mostly except from the MT data the passive
tomography data has been used. The resulted model has been compared with the results
from the MT data. A description of both the passive tomography interpretation and the MT
data interpretation are presented from the work package (WP5) leader (UTHESS.GPL).
D7.Dissemination of results to Industry is already obtained inside the consortium since one of the
partners is a private oil company (ENTOIL). However, a web site has also been created for
further dissemination of the results to the industry. A presentation of the processing tool has
been accomplished into the electromagnetic group of the University of Uppsala, Sweden.
Finally two presentations of the interpretation results of the project following two extended
abstract has been held in the EAGE conference in 2002. A report presenting the form of the
web site, the two publications and the economic comparison between the methods applied an
previous methods is available from the work package (WP6) leader (UTHESS.GPL).
Final Publishable Report
21
CEHEG/ENK6-2000-00100
6
Scientific and technical performance
6.1 Comparison of initially planned activities and actual work accomplished.
The first deviation that should be commented is the increase of man-months (MM) necessary for
finalising Work Package 2, i.e. “Data Acquisition”. There was an increase of 1.5 MM for
UTHESS.GPL and that was due to the fact that it was necessary to spend more MM to check
another possible area of study, i.e. the area of Delvinaki, since more geological and geophysical
information was available for that area compared to the initially selected area of study, i.e. Filliates.
An increase of 1.2 MM for GEOSYST during the fieldwork was due to the successful attempt of
increasing the number of MT soundings per day. A maximum number of seven stations per day
were achieved although the scheduled set-up was with four stations per day. In order to accomplish
such a number GEOSYST provided more field engineers than we initially planned.
The second deviation is the less MM already spent for Work Package 4, i.e. “Modelling Code”. At
first the increase on the MM of UTHESS.GPL for subtask 4.1, “Analysis”, was due to the fact that
the theoretical background necessary for the analysis of the inversion algorithm was higher than
expected for the personnel of the University of Thessaloniki, so more work had to be done on this.
On the other hand GEOSYST has been working in this subject before the start of the project so
there was a decrease on their MM. The personnel of GEOSYST helped very much so that their
‘know how’ about the specific task were transferred to the personnel of UTHESS.GPL.
A decrease in MM is present for Work Package 5, i.e. “Interpretation” for both partners,
UTHESS.GPL and GEOSYST. That was, for UTHESS.GPL, due to the fact that there was less
geophysical data of good quality than expected (only MT and Passive Tomography Data were
usable). GEOSYST on the other hand was very much acquainted with working in such problems of
interpreting MT data that needed less MM to work on the Work Package.
6.2
State of the art review: comparison of the project current work with the competing
technologies and assessment of its implications on the relevance of the project.
The principal competing technology in energy exploration is the seismic reflection method. Figures
from the Society of Exploration Geophysicists indicate that about 290 seismic crews are active
worldwide, compared with about 5 MT in the world outside of China. There practical reasons for this
imbalance:
a)
The seismic method is capable of very high resolution at depths of exploration interest.
b)
Massive investments (of the order of 0.5 to 1 billion US dollars over the last decade) by oil
companies and by service companies have driven instrumentation and processing to very high
levels.
c)
The MT method is not capable of high resolution; although this is offset by the ability to
acquire useful data where seismic fails.
d)
MT is little known, investments in interpretation have been minimal (perhaps one million
dollars over the last decade), and the number of capable practitioners is at most 10 (world
wide).
In the case of the current project, seismic surveys have delivered little usable data due to the
presence of outcropping high-velocity and often steeply dipping carbonates and anhydrite. By way
of contrast, these rocks are virtually transparent to MT and imaging below them has presented no
insurmountable problems.
This EU-funded program potentially has two significant implications with regard to seismic: the
availability of a significant body of MT data, which can be published, and the advances in
Final Publishable Report
22
CEHEG/ENK6-2000-00100
interpretation. A limiting factor in the past, for example, has been that MT surveys carried out for
oil companies have remained proprietary, and without interpreted demonstration data it is difficult
to demonstrate the utility of the method.
In this project, mostly the personnel of GEOSYST have worked out optimisation of the modelling
code. The personnel of UTHESS.GPL together with ENTOIL and GEOSYST worked in the
interpretation of the geophysical data (seismic, passive tomography and well stratigraphy) that are
already available to the consortium. The final 3-D model has been compared with the other
geophysical data and an interpretation was possible presenting the possible geology of the area of study
until a certain depth.
Final Publishable Report
23
CEHEG/ENK6-2000-00100
7
Management and co-ordination aspects
At the beginning of the project a kick-off meeting was hosted by ENTOIL in its office in Athens
where WP1 and WP4 were discussed between UTHESS.GPL and GEOSYST, and WP2 between all
partners. Discussion for WP1 and WP2 was based in a QA framework, just to confirm the way each
of the partners will be involved, as it was already scheduled in the Technical Annex of the CPF.
Discussion for WP2 was very productive since all aspects of the fieldwork area, and the availability
of geophysical/geological data was evaluated. The meeting was held in a constructive and cooperative style.
Dr. Savvaidis from UTHESS.GPL visited the offices of GEOSYST in Milan during February 2001,
where he was hosted for four days. In a very warm welcome atmosphere, results from WP1 were
discussed and the field planning along with the task 4.1 of WP4 (analysis of modelling code). An
overview of the maps, the geophysical and the geological data were presented by UTHESS.GPL
and the final decision for the fieldwork area was made. ENTOIL was then informed and agreed for
the proposed area for fieldwork.
During the fieldwork there was a very warm cooperation between the personnel of UTHESS.GPL
and the field engineers of GEOSYST. An every day communication (by e-mail and by phone if
necessary) between the scientific persons in charge of the project from each partner was vital and
enforced the quick and excellent data acquisition. Dr Savvaidis from UTHES.GPL and Mr Zalai
from GEOSYST were the two field leaders from each of the partners involved in the fieldwork.
Professor Tsokas as a coordinator was present in the mobilization and demobilization of the field
crews.
A one-day meeting of Dr Karageorgi and Mr Watts were held in Milano in the office of GEOSYS at
the end of July 2001 with the specific objective of evaluating the preliminary MT interpretation in
the light of existing data and geological concepts. For ENTOIL this preceded the spudding of a
deep exploration well in the area.
The personnel of UTHESS.GPL mostly involved in the interpretation of the geophysical data were
met in ENTOIL offices in Athens in order to finalise the resulted models. Also, different scientific
teams from UTHESS.GPL visited ENTOIL offices in Athens for discussing the results of the
project and get acquainted with the decision-making procedures of the well planning. That was part
of the transfer of the know how of the oil company to the personnel of the UTHES.GPL.
The communication between partners was mainly through Internet (all three partners have a direct
Internet Connection). Also, via the exchange of maps that exist only in paper form and CD’s
including digital geological and geophysical data.
Based on Table 3 there was an increase of man-months (MM) necessary for finalising Work Package
2, i.e “Data Acquisition”. There was an increase of 1.5 MM for UTHESS.GPL and 1.2 MM for
GEOSYST that was due to the fact that it was necessary to check another possible area of study, i.e.
the area of Delvinaki from the personnel of UTHESS.GPL and due to more field engineers that
GEOSYST provided in the fieldwork to increase the rate of MT stations per day. An increase on the
MM of UTHESS.GPL for subtask 4.1 was due to the fact that the theoretical background necessary for
the analysis of the inversion algorithm was higher than expected for the personnel of the University of
Thessaloniki, so more work had to be done on this. On the other hand GEOSYST has been working in
this subject before the start of the project so there was a decrease on their MM for the same subtask.
Finally, work package 4 finished with less MM from GEOSYST on the second year, as expected.
Also, GEOSYST’s interference in work package 5 (Interpretation) was with less MM than planned
since its personnel was already acquainted with such problems and it was very easy for them to
understand the geology of the area. Also, for UTHESS.GPL it was not necessary to spend all the
planned MM for work package 5 since the interpretation has been done only in MT and passive
Final Publishable Report
24
CEHEG/ENK6-2000-00100
tomography data. The interference of the research group from ENTOIL was vital in that stage since
their know how from different geological problems in other parts of the world was very important.
The increase of MM for the first year of UTHESS.GPL for work packages 1 to 4 has been balanced
with work package 5 were there was a decrease in the man power necessary to interpret the
geophysical data.
A bar chart with the original work-plan per task together with an update is given at Table 4. The only
deviation noted is the increase of the time necessary for the implementation of the modelling code,
which needed to be extended since some time was spent to transfer the know-how of GEOSYST in the
analysis of the modelling code to UTHESS.GPL.
During the first year of the project the UTHES.GPL released a sub-contract for the rental of
geophysical equipment necessary for the fieldwork. This subcontract due to a small increase in market
prices compared to the prices stated in the “Justification of Costs” during the CPF, exceeded the
planned cost (Table 5). The initial cost was 53100 EURO but the final sub-contract was at a price of
60000 EURO. To balance this deviation we had to transfer money from other three categories (labour,
travel and consumables) of the eligible cost table to sub-contracting. This was established with the
permission of the scientific manager, Mr. John Garnish.
Also an increase in the actual personnel cost was noted for the UTHESS.GPL since more man-months
were necessary for the establishment of different tasks of the project. A decrease of the travel and
subsistence costs has been noticed compared to the planned numbers. This amount of money together
with the decrease in consumables balanced the increase on the categories of sub-contracting and
labour.
Since SHELL has taken over ENTOIL a new contact person has been appointed, i.e. U.J. Seeman.
Partner
Contact Person
Contact Information
UTHESS.GPL
Prof. Gregory Tsokas Addrress: Geophysical Laboratory, School of Geology,
University
of
Thessaloniki,
54006,
Thessaloniki, GREECE
Phone:
+302310998507
FAX:
+302310998528
e-mail:
gtsokas@geo.auth.gr
GEOSYST
Mr Michael Watts
Addrress: Geosystem srl, Viale Abruzzi 17, 20131,
Milan, ITALY
Phone:
+390229404727
FAX:
+390229404577
e-mail:
dwatts@geosystem.net
ENTOIL/SHELL Mr Uli Seemann
Addrress: SRM B.V. Visseringlaan 26, 2288 ER
Rijswijk, Netherlands
Phone:
+31703113313
FAX:
+31703113312
e-mail:
Uli.U.J.Seemann@shell.com
Final Publishable Report
25
CEHEG/ENK6-2000-00100
Table 3 : Man Power and Progress Follow-up Table
- - - - - - - - - Man-Month - - - - - - - - - - Partner
(Name/
abbrev.)
Task/Subtask
Planned efforts at
start of period (MM)
Year 1 Year 2 Total
Year 1 Year 2
- - - - - - - Technical Progress % - - - - - - Deviation
(MM)
Actual
effort
(MM)
Total
Totals
d1
d1-d
0.5
(N°/title)
A (UTHESS.GPL)
B(GEOSYST)
C(ENTOIL)
a
A
B
1. Processing Code
1.1 Analysis
b
b1
Planned
(%)
Assessed*
(%)
Year Year Year
1
2
1
Year
2
100%
100%
100%
100%
100%
100%
100%
100%
d
a1
1.0
1
1.5
1.5
2.0
2
2
2
A
1.0
1
1
1
100%
100%
100%
100%
B
2.0
2
3
3
1
100%
100%
100%
100%
C
Total
6.0
6
7.5
7.5
1.5
100%
100%
100%
100%
A
2.0
2
3
3
1
100%
B
0.5
0.5
0.5
0.5
C
0.2
0.2
0.1
0.1
A
2.0
2
2.5
B
1.8
1.8
3
Deviation
(%)
Year
(now)
Comments on
major
deviations
and/or
modifications
of planned
efforts.
C
1.2 Implementation
2. Data Acquisition
2.1 Planning
2.2 Fieldwork
100%
100%
100%
100%
100%
-0.1
100%
100%
100%
100%
2.5
0.5
100%
100%
100%
100%
3
1.2
100%
100%
100%
100%
0.1
9.2
0.1
2.7
100%
100%
100%
100%
C
Total
6.5
6.5
0.1
9.2
A
2.0
2
2.5
2.5
0.5
100%
100%
B
2.0
2
2.5
2.5
0.5
100%
100%
C
Total
4.0
4
5
5
1
100%
100%
100%
A
1.0
1
2
2
1
100%
100%
100%
B
3.0
3
2
2
-1
100%
100%
100%
100%
Check another
area for fieldwork. Increase
the production
rate using more
field engineers.
3. Processing Optimization
100%
4. Modelling Code
4.1 Analysis
C
4.2 Implementation
A
2.0
2
0.5
1.5
2
100%
100%
25%
100%
B
4.0
4
0.5
3
3.5
-0.5
100%
100%
14%
100%
C
Total
10.0
10
5
4.5
9.5
-0.5
100%
100%
53%
100%
GEOSYST well
acquainted with
modelling code
UTHESS.GPL
not acquainted
5. Interpretation
A
8.8
8.8
6.25
6.25
-2.55
100%
100%
B
2
2
1
1
-1
100%
100%
C
Total
0.8
11.6
0.8
11.6
0.8
8.05
0.8
8.05
-3.55
100%
100%
100%
100%
Less data to be
interpreted Good know how
6. Reporting
6.1 Results
6.2 Exploitation
TOTALS
A
2
2
2
2
100%
B
0.5
0.5
0.5
0.5
100%
100%
C
0.5
0.5
0.5
0.5
100%
100%
A
1.9
1.9
2
2
100%
100%
B
0.5
0.5
0.5
0.5
100%
100%
C
Total
5.4
5.4
5.5
5.5
0.1
100%
100%
0.1
A
11.0
12.7
23.7
13.0
11.8
24.8
1.05
46%
100%
53%
100%
B
15.3
3.0
18.3
13.5
5.0
18.5
0.2
84%
100%
73%
100%
C
TOTAL
0.2
26.5
1.3
17.0
1.5
43.5
0.2
26.7
1.3
18.1
1.5
44.8
1.25
13%
61%
100%
100%
13%
60%
100%
100%
*) Please note that the actual technical progress percentage and the updated remaining efforts must reflect the physically assessed status of the work.
Final Publishable Report
26
CEHEG/ENK6-2000-00100
Table 4 : WORKPACKAGE / MANPOWER BARCHART
Duration / critical path
:see milestone list
1
1st year
2nd year
Workpackage 1: Processing Code
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
Workpackage
descriptions
1
Task 1.1 Analysis
Task 1.2 Implementation
Workpackage 2: Data Acquisition
1
Task 2.1 Planning
Task 2.2 Fieldwork
Workpackage 3: Processing Optimization
1
Workpackage 4: Modeling Code
1
Task 4.1 Analysis
Task 4.2 Implementation
1
Workpackage 5: Interpretation
Workpackage 6: Reporting
Task 6.1 Results
Task 6.2 Exploitation
6m
Mid-term
18 m
Final
Reporting Periods
Work-package/Manpower Bar-chart: Blue, double-headed arrows show the original plan per work-package and black lines per task. Red, black
outlined, double headed arrows, show the extended work-package and the dashed red line the relevant task.
Final Publishable Report
27
CEHEG/ENK6-2000-00100
Table 5 : Budget Follow-up Table
PARTNER
P1(UTHESS.GPL)
P2(GEOSYST)
Cost Category
ACTUAL COSTS
(EUR)
BUDGET
(EUR)
Total Pct. Spent
(%)
Remaining
Budget
(EUR)
Year 1
Year 2
Year 3
Year 4
Total
Year 1
Year 2
Year 3
Year 4
e
a1
b1
c1
d1
e1
a1/e
a1+b1/e
a1+b1+c1/e
a1+b1+
c1+d1/e
Labour
59120
34482.76
26763.7
61246.46
58%
104%
-2126.46
Overheads
26424
17216.19
7864.86
25081.05
65%
95%
1342.95
Labour+Overheads
85544
51698.95
34628.56
86327.51
60%
101%
-783.51
Travel
55000
36295.7
11944.97
48240.67
66%
88%
6759.33
Durable Eqmt.
5000
4958.67
4958.67
99%
99%
41.33
Consumables
5000
1991.01
2606.66
40%
52%
2393.34
External Assistance
53100
60059.35
60059.35
113%
113%
-6959.35
Computing
8000
8352.81
8352.81
104%
104%
-352.81
Car Rental
Total
211644
163356.49
47189.18
210545.67
77%
99%
1098.33
Labour
135765
118070
17976
136046
87%
100%
-281
Overheads
61094.25
53131.5
8089.2
61220.7
87%
100%
-126.45
Labour+Overheads
196859.25
171201.5
26065.2
197266.7
87%
100%
-407.45
4000
2850
750
3600
71%
90%
400
200859.25
174051.5
26815.2
200866.7
87%
100%
-7.45
45914
7024.05
38915
45939.05
15%
100%
-25.05
Labour+Overheads
45914
7024.05
38915
45939.05
15%
100%
-25.05
Travel
6000
6000
6000
44915
51939.05
14%
100%
-25.05
Travel
615.65
Comments on major deviations from budget.
e-e1
Durable Eqmt.
Consumables
External Assistance
Computing
Car Rental
Total
P3(ENTOIL)
Labour
Overheads
100%
Durable Eqmt.
Consumables
External Assistance
Computing
Car Rental
Total
Etc……………..
Etc……………..
TOTAL
Labour
51914
7024.05
240799
159576.81
83654.7
243231.51
66%
101%
-2432.51
Overheads
87518.25
70347.69
15954.06
86301.75
80%
99%
1216.5
Labour+Overheads
328317.25
229924.5
99608.76
329533.26
70%
100%
-1216.01
Travel
65000
39145.7
18694.97
57840.67
60%
89%
7159.33
Durable Eqmt.
5000
4958.67
4958.67
99%
99%
41.33
Decrease since it was not necessary (manage to finish
fieldwork in les time)
Consumables
5000
1991.01
2606.66
40%
52%
2393.34
Decrease for balancing other expenses (Sub-Contracting)
External Assistance
53100
60059.35
60059.35
113%
113%
-6959.35
Increase in rental of equipment due to change of market prices
Computing
8000
8352.81
8352.81
104%
104%
-352.81
Car Rental
Total
464417.25
344432.04
463351.42
74%
100%
1065.83
Final Publishable Report
615.65
Increase of labour since it was necessary for UTHESS.GPL
118919.38
28
CEHEG/ENK6-2000-00100
8
Results and Conclusions
The work already accomplished is a state of the art research and development in the subject of the cost
effective hydrocarbon exploration using the magnetotelluric method. A robust multi-station processing
code has been implemented and the code has been tested onto previously collected test data and the
data acquired through the work of this project. The later data set helped very much in optimising the
processing code. This data set of 99 MT stations could be used as a reference since it consists of evenly
spaced soundings that follow a grid of 9 rows by 11 columns. The processed data gave as a result the
transfer functions connecting the electric and magnetic field at each site as a result of the processing
code implemented through this project. It is the first time in Europe that such a big amount of MT
soundings exist in one area following a grid of approximately 100 square kilometres, that allow two
and three-dimensional modelling.
Two and three-dimensional modelling codes for magnetotelluric data have been implemented through
the project based on the previous know how of one of the partners. A 2D magnetotelluric sharp
boundary inversion algorithm has been developed that inverts MT data for a 2D model consisting of
discrete interfaces and the resistivities of the layers between those interfaces. Also a 3D
magnetotelluric inversion algorithm that is based on previous work of the personnel of one of the
partners on 2D MT inversion and 3D MT forward modelling has been developed. With this code the
time to the first 3D model has been minimised to some hours of computations. In this way the end-user
will move to the research/productive wells faster than on the conventional way, where a lot of time for
processing and modelling was necessary prior to the first drill.
Different geophysical and geological models were the results of the interpretation of different data sets.
Mostly the data acquired through this project were used using as constrains the well information
provided by the oil company. The three dimensional model provided to the consortium map very well
the main different geological formations of the area of study.
The application of the magnetotelluric method in hydrocarbon exploration is vital since we managed in
a very cost effective way using an environment friendly geophysical method. That is very important in
both hydrocarbon and geothermal exploration since it might change the way of working for the
Exploration Teams. The major impact could be in the oil industry since they will not have to pay
enormous amount of money for the exploration and in this way they could apply more productive
wells.
Finally, we can conclude to the completion of a tool that manages for the first time to compete the
seismic method in all aspects. A software tool that cover processing and modelling of MT data has
been implemented. The MT method has been applied in an urban area of difficult geology and
topography and produced through the implemented software a three dimensional image of the earth. In
that way the time to the first 3D model of the area of study has been eliminated with a cost effective
way.
Final Publishable Report
29
CEHEG/ENK6-2000-00100
9
Acknowledgments
First of all we would like to thank all the “Scientific and Financial Managers” of the project,
especially Mr John Garnish who was very helpful at the beginning of the project during the Contract
Preparation Form procedure and the first year of the project, and Mr Jeroen Schuppers who guided us
for the final outcome of the project and showed extreme patience for the final reports.
We would like to thank all the personnel of Geosystem Srl and ENTOIL Ltd that were not involved in
the project and were very helpful and welcome in their co-operation and through the meetings that they
hosted.
Also we would like to thank the personnel of SHELL, Mr John Finlay and Mr Uli Seemann where very
much helpful and co-operative although they were involved in the end of the project since SHELL took
over ENTOIL at the end of the project.
The scientific personnel involved in the project and the filed work from the Aristotle University of
Thessaloniki would like to thank both the personnel from Geosystem Srl and ENTOIL Ltd for their
guidance and transfer of knowledge though the project. It is considered of great success that two
geophysicists from the University of Thessaloniki are part time working with Geosystem Srl at the end
of the project.
Finally we would like to thank the Geophysical Laboratory of the Aristotle University of Thessaloniki
who provided its infrastructure, especially their computing facilities, without which it would not be
possible to finish this project at the proposed schedule and with such benchmarks in computation time
and power.
Final Publishable Report
30
CEHEG/ENK6-2000-00100
10 Glossary
AC: Alternate Current
CD: Compact Disc
CPF: Contract Preparation Forms
CPU: Central Processing Unit
DC: Direct Current
DFT: Discrete Fourier Transform
EGS: European Geophysical Society
EM: Electromagnetic
ENTOIL: Enterprise Oil Ltd
FFT: Fast Fourier Transform
GEOSYST: Geosystem Srl
GIS: Geographical Information System
GUI: Graphical User Interface
HZ: Hertz
KM: Kilometers
MM: Man-months
MT: Magnetotelluric
N/A: Not Available
OS: Operating System
PC: Personal Computer
QA: Question Answer
TDEM: Time-Domain Electromagnetic
UTHESS.GPL: University of Thessaloniki, Geophysical Laboratory
WP: Work Package
Final Publishable Report
31
CEHEG/ENK6-2000-00100
11 References
Chave, A.D., and Thomson,
D.J.
1989
Some comments on magnetotelluric response function
estimation: J.Geophys.Res., 94, 14215-14225
Egbert, G.D.
1997
Robust multiple-station magnetotelluric processing:
Geophysical Journal International, 130 475-496.
Jones, A.G., Chave, A.D.,
Auld, D., Bahr, K. and Egbert,
G.
1989
A comparison of techniques for magnetotelluric response
function estimation: J. Geophys. Res., 94, 14,201-14,213,
1989
Larsen, J., Mackie, R.L.,
Manzella, A., Fordelisi, A., and
Rieven, S.
1996
Robust smooth magnetotelluric transfer functions:
Geophysical Journal International, 124 801-819.
Mackie et al.
1994
Rodi and Mackie
2001
Nonlinear conjugate gradients algorithm for
magnetotelluric inversion, Geophysics, 66, 174-187.
Trad, D., and Travassos, J.
2000
Wavelet filtering of magnetotelluric data: Geophysics
65(2) 482-491.
Wight, D.E., and Bostick, F.X.
1981
Cascade decimation-a technique for real time estimation
of power spectra: Proc. IEEE ICASSP 626-629
Final Publishable Report
Three-dimensional electromagnetic modelling using
finite difference equations: The magnetotelluric
example, Radio Science, 29, 923-935.
32
2-D
CEHEG/ENK6-2000-00100
Final Publishable Report
33
CEHEG/ENK6-2000-00100
12 Annex (Individual Reports for each Work Package as Provided by the Work
Package Leader)
Final Publishable Report
34
CEHEG/ENK6-2000-00100